DE10308881B3 - Control arrangement and method for checking the function of such a control arrangement for occupant protection means in a motor vehicle - Google Patents

Abstract

The present invention is based on the idea of not multiplying the output signal (a) of an acceleration sensor (19) with a correction value (k¶w¶) with regard to a weighted sum (SIGMA¶g¶), but by means of a weighting means (16 ) to modify a test signal (t) in such a way that an already weighted output signal (a¶g¶) can be generated, so that the safing algorithm of an evaluation device can be used directly during a test operation, which means that this is advantageous even with regard to its functionality is verifiable. DOLLAR A The present invention is particularly suitable for occupant protection systems of a modern motor vehicle.

Description

The invention relates to a control arrangement as well a method for checking the function of such Tax arrangement for Passenger protection means in a motor vehicle.

Occupant protection means such as airbags, Belt tensioners or the like represent an important safety component of today's motor vehicles. Although they are not required by law, however, almost every car currently owns somewhere in the world is produced, at least one front airbag for the driver.

Many automobiles also have the front airbag already has other airbags, esp. Side airbags, head airbags, knee airbags or the like for drivers, Passenger or other occupants. Each of these airbag systems uses several Sensors arranged at various locations on the vehicle body and what is the deceleration (negative Acceleration). The usual used systems work with driver, front passenger and side airbags with acceleration sensors arranged in or adjacent to a control arrangement. The control arrangement is located at a central point in the motor vehicle, for example under the driver's seat or in the vehicle tunnel. It is therefore often referred to as the central module. To detect a side impact are at least one acceleration sensor on each side of the vehicle - or more advanced at least one pressure sensor - provided, which, due to their decentralized arrangement, are often called satellites be designated. The respective sensor is generally relatively high Requirements because it is the first component of an occupant protection system, at which the crash information arrives. He has to delay the fast Motor vehicle in a reliable and implement exact electrical signal (a).

The most common methods of acceleration measurement counts the Measuring the effect of a force F resulting from a seismic Mass m acting acceleration g results. This force creates mechanical stresses and a change in position of the seismic mass. The tensions can based on the piezoresistive (or piezoelectric) properties of the material used can be determined. Changes in location are usually under Measured using a variable capacity. The piezoresistive Effect in semiconductors is widely used in pressure sensors, while for acceleration sensors the capacitive measuring principle in a variety of technical applications is preferred. With this design, surface microfabrication can be used realize very small sensor structures and thus cost-effective solutions. The capacitive built sensors are also less prone to temperature fluctuations and offer a great Operating temperature range. That is why they are based on the field of occupant protection systems in addition to the acceleration sensors, the pressure sensors used are also predominantly used on this principle.

The actual sensor element, the so-called "G-cell" (Gcell) is special a mechanical construction made of semiconductor materials. For example, it consists of two fixed plates with an intervening moving plate that holds the seismic mass represents. If the G-cell is subjected to acceleration, it moves the middle tile its rest position. If the middle tile moves, elevated its distance to one of the solid platelets to the same extent as itself the distance to the other is reduced. The change in distance is a Measure of the acceleration. The one for hanging of the middle plate carrier act as feathers; a possibly compressed fluid between the platelets, for example a special gas or just air, dampens the Move. If this is undesirable is known to provide a vacuum. A G cell senses in usually along a sensitivity axis. By skillful A mass can also be used for two axes, which reduces costs. One then speaks of so-called x-y-Gcells or X-Y sensors. In electrical terms, the plates form the G cell a coupled capacitor pair. If the middle Tiles as a result the acceleration moves along the sensitivity axis changes the distance between the plates, which also increases the capacity each of the two capacitors changes. The same applies to G cells with e.g. several interdigitated elements.

The G-cells manufactured by microfabrication have very small dimensions. The seismic mass sometimes weighs only a few hundred picograms (1 picogram = 10 ^–12 grams). When an acceleration of 100 g is applied, the movable plate or the finger changes its position by less than 400 nm (nanometers). In order to achieve a measurement resolution of 1 g, a change in capacitance ΔC of less than 1 femtofarad (10 ^-15 F) must be recognized. In order to be able to measure such a small capacitance, a dedicated control circuit is required in the acceleration sensor, which converts the capacitance into a usable output signal (a).

The output signal (a) from the sensor becomes an evaluation unit, which has at least one, nowadays usually includes several microcontrollers, supplied by followed by one executes a crash discrimination algorithm to distinguish between an actual one Differentiate impact and normal dynamic vehicle behavior and possibly a trigger signal for the Retention means generated.

The trigger signal is currently often only dependent on a so-called switching signal, which in the simplest case, for example can come from a mechanical acceleration switch. In a Nowadays, however, a large number of acceleration sensor arrangements are adopted the acceleration sensors themselves do this task. Such so-called Safing sensors ensure after execution the so-called saving algorithm, for the regularly nowadays own microcontroller is provided in the evaluation unit for release or otherwise for a prevention of triggering the restraint, if the acceleration sensor or the evaluation device, i.e. the algorithms executed in the microcontrollers are faulty work and would therefore deliver a faulty trigger signal.

From the EP 1 149 004 - The disclosure of which is expressly included in full - a method and a device for checking the function of a control arrangement for occupant protection means in a motor vehicle are known, in which a weighted sum is formed from the output signals of the acceleration sensors for checking the plausibility of the signals, by at least the output signal of an acceleration sensor is multiplied by a correction value. Such a function check advantageously allows statements about the functionality of the acceleration sensors, their signal powers, levels, or the like. However, it is not possible to say to what extent the safing algorithm itself works reliably because it is not used during the test mode.

The object of the invention is a improved method for checking the function of a system of several Acceleration sensors of a control arrangement for occupant protection means in a motor vehicle to deliver. In particular, in addition to the sensors, the functionality should be of the safing algorithm itself can be checked his.

This object is achieved by a Circuit arrangement with the features according to claim 1 and by a procedure for checking functionality the features according to claim 7 solved.

The present invention is based on the idea of not first multiplying the output signal (a) of an acceleration sensor with regard to a weighted sum (Σ _g ) by a correction value (k _w ), but of changing a test signal (t) by means of a weighting means such that an already weighted output signal (a _g ) can be generated, so that the safing algorithm of an evaluation device can be used directly during a test operation, as a result of which it can be advantageously checked even with regard to its functionality.

Advantageous training and further education, which can be used individually or in combination with each other, are in the dependent claims specified.

Further advantages of the invention and their further developments are described below using exemplary embodiments and closer to the drawing explained.

It shows schematically:

1 the typical structure of an occupant protection system in a motor vehicle;

2 an acceleration sensor comprising a seismic mass which can be deflected along a sensitivity axis;

3 an acceleration sensor comprising a seismic mass which can be deflected along two sensitivity axes;

4 the acceleration sensor 2 in operative connection with a weighting agent;

5 an arrangement with an XY acceleration sensor and a further acceleration sensor;

6 an alternative sensor alignment according to the arrangement 5 ; and

7 a star-shaped arrangement of three acceleration sensors.

The same elements or signals are all figures identified by the same reference numerals.

1 shows the typical structure of an occupant protection system in a motor vehicle 1 , At as central a location as possible in the motor vehicle 1 there is a control arrangement 2 , This includes an evaluation unit 3 for example in the form of at least one microcontroller. In the tax arrangement 2 or there is a sensor field adjacent to it 5 in which suitable sensors 17 . 18 . 19 . 20 for measuring accelerations, for example an acceleration g _x along a sensitivity axis in the x direction or g _y along a sensitivity axis in the y direction. The sensitivity axes u, v, w, x, y of the sensors 17 . 18 . 19 . 20 span a level, which after installation of the control arrangement 2 in a motor vehicle 1 is essentially parallel to a plane defined by the vehicle longitudinal axis AA 'and the vehicle transverse axis BB'. More sensors 6 , in particular for the detection of a side impact, are for the preferably capacitive measurement of lateral accelerations, for example an acceleration g _r from the right or g _l from the left, at a decentralized location in the side of the motor vehicle 1 arranged. As sensors 6 For Acceleration sensors are typically used for side installation, but more recently pressure sensors have also been increasingly used. The respective output signals a of the sensors are from those in the evaluation unit 3 arranged microcontrollers, one of which then executes a crash differentiation algorithm in order to differentiate between an actual impact and the normal dynamic vehicle behavior. A microcontroller of the evaluation unit, which is preferably independent of the crash processing 3 also performs a continuous and / or cyclical diagnosis of the system using a safing routine to ensure that it is working properly and is available in the event of an accident. The one in the central sensor field 5 arranged sensors 17 . 18 . 19 . 20 Just like the 6 on the side, they have to be extremely reliable so that they 3 do not send false signals a, which could lead to undesired activation of the restraint. The driver is therefore informed of any malfunction, for example by an airbag warning lamp on the instrument panel (not shown), and the restraint function is blocked until the fault has been rectified. If the airbags have to be deployed in the event of an impact, the evaluation unit activates 3 an ignition switch 4 so that current flows through the ignition circuits of the igniter for the driver's front airbag 7 , the igniter for the front passenger airbag 8th , the igniter for the side airbag 9 , the detonator for the belt tensioner 10 , or the like flows more, with which the belt tensioners are activated and the gas generation reaction is triggered within the inflation modules.

2 shows the functional principle of a capacitive acceleration sensor 17 . 18 or 19 comprising a sensor element 11 - hereinafter also as a G cell 11 designated - in which a seismic mass 12 is arranged to be deflectable along a sensitivity axis w. The mode of operation is based, for example, on micromechanical double-plate capacitors, which in the present case is finger-shaped. Between two outer rigid plates 13 is a middle plate 14 on the movably suspended seismic mass 12 fixed. The mass shifts during acceleration 12 so that the capacity changes. It is also known to arrange a plurality of interdigitated elements one behind the other. By skilful construction, a mass can 12 can also be used for two sensitivity axes (x, y), which advantageously reduces costs.

3 shows such a so-called capacitive XY acceleration sensor comprising a sensor cell 11 in which a seismic mass 12 is arranged to be deflectable along two sensitivity axes x and y. As in the sensor 2 is the G cell 11 a dedicated control circuit 15 downstream, which converts the capacity into a usable output signal (a).

In normal, ie crash, operation of the evaluation device 3 all output signals a _u , a _v , a _w and a _w , a _x , a _{y of} the sensors 17 . 18 . 19 respectively. 19 . 20 checked for plausibility by means of a safing algorithm by forming a weighted sum Σ _g from the output signals a _u , a _v , a _w or a _w , a _x , a _y . For example, in parallel to this, the output signals a _u , a _v , a _w or a _w , a _x , a _{y are} evaluated by means of a crash differentiation algorithm, the restraint means only being released if the plausibility has been established. According to the invention, preferably at least one of the output signals a _u , a _v , a _w or a _w , a _x , a _{y is} compared with a threshold value SW, so that only when the threshold value SW is exceeded by at least one of the output signals a _u , a _v , a _w or a _w , a _x , a _y the safing algorithm is released.

To when starting and / or during the operation of the motor vehicle 1 the correct functioning of the control arrangement 2 To be able to check, it is advantageous to check the acceleration sensors 17 . 18 . 19 respectively. 19 . 20 to be able to undergo a so-called self-test. For this purpose, the control arrangement sends in test mode 2 first the evaluation unit 3 , for example a microcontroller, a test signal t at at least two acceleration sensors 17 . 18 . 19 respectively. 19 . 20 , so as to generate output signals a _u , a _v , a _w or a _w , a _x , a _y , which the function check of the sensors 17 . 18 . 19 respectively. 19 . 20 serve. In the case of a so-called physical test of the sensor (s) 17 . 18 . 19 . 20 causes the test signal t to deflect the seismic mass 12 along the sensitivity axis u, v, w, x, y. The capacity change in the G cell 11 is in one of the G cells 11 subordinate control circuit 15 of the sensor 17 . 18 . 19 . 20 recognized and uses this knowledge for the microcontrollers of the evaluation unit 3 usable output signal a. Such a physical test advantageously allows both a statement about the functionality of the G cell 11 as well as the functionality of the control circuit 15 of the acceleration sensor 17 . 18 . 19 . 20 ,

It is also known only the control circuit 15 of the acceleration sensor 17 . 18 . 19 . 20 to supply a test signal t by means of which a usable output signal a is generated or simulated. In such a purely electronic test, no statements can be made about the functionality of the G cell. The only statement about the electronic functionality of the control circuit 15 of the sensor 17 . 18 . 19 . 20 may be sufficient in some cases, especially if other mechanisms for detecting a defective G cell are provided, such as measuring displacements or fluctuations in the offset voltage of the sensor insofar as these are characteristic of a defective G cell.

4 shows an acceleration sensor according to the invention 17 . 18 . 19 or 20 , This is behind the acceleration sensor 2 modified to the extent that it is additionally weighted 16 is in operative connection. The weight means 16 can be part of the sensor 17 . 18 . 19 . 20 and / or part of the control arrangement 2 , especially part of the evaluation device 3 his. This advantageously allows one of the evaluation unit 3 to change the incoming test signal t such that a physical test results in a weighted deflection of the seismic mass 12 done itself. In this case, especially so-called test fingers or the like which are specially formed in the sensor can be considered as weighting means. During a self-test, the test fingers alone experience a deflection. Arranged on the same seismic mass, their deflection also causes a deflection of the fingers provided for the crash sensing. Depending on the desired degree of importance of an output signal, more or fewer test fingers are deflected. If - to put it bluntly - instead of the standard ten test fingers, for example, fourteen or even only seven test fingers are deflected, this results in a correspondingly increased or reduced deflection of the fingers intended for crash detection and a correspondingly weighted output signal. But an electronic test is also possible by using the weighting device 16 a weakened or amplified, ie weighted test signal t _g , the control circuit 15 supplies. As a weighting device 16 Voltage-reducing electronic components such as resistors or voltage-increasing electronic components such as, for example, an electronic charge pump or the like are therefore suitable: one such, a weighting means 16 comprehensive acceleration sensor 17 . 18 . 19 . 20 allows unweighted acceleration sensors to output a weighted output signal a _g with the same self-test signal t received on all sides.

The present invention now makes use of this consideration by providing at least one sensor 19 provides which outputs a weighted output signal a _g . Of course, depending on the arrangement of the acceleration _sensors , the generation of several weighted output _signals a _ug , a _vg , a _wg , a _xg , a _{yg is} sometimes expedient. Various arrangements preferred according to the invention show 5 to 7 ,

5 shows the sensor field 5 a control arrangement 2 comprising three acceleration sensors 17 . 18 . 19 with one G cell each 11 , which allows the sensation of an acceleration g along a sensitivity axis w, x, y, which are each oriented differently. sensor 17 serves the sensation in X-; sensor 18 in the Y direction. sensor 19 is arranged inclined to them. In the downstream evaluation device 3 can already operate in crash mode due to only two of the three acceleration sensors 17 . 18 the direction and strength of one on the vehicle 1 acting acceleration g can be determined. The signal a _{w from} the third acceleration sensor 19 is used to check one of the two calculated quantities, the direction or the strength of the vehicle 1 acting acceleration, used. The third sensor 19 thus takes over the function of a safing sensor and can trigger the restraint in this way 7 . 8th . 9 . 10 Prevent at least indirectly if the value provided by it deviates significantly from a value calculated from the signals a _x and a _{y of} the other two sensors.

The sensors 17 and 18 are unmodified and customary in the example, ie they generate the test arrangement t after receiving a test signal in test operation of the control arrangement 2 an unweighted output signal a _x or a _y . sensor 19 is a constructed according to the invention, with a weighting means 16 active acceleration sensor 19 whose sensitivity axis w is the sensitivity axis of the X or Y sensor 18 respectively. 17 , for example by 45 ° to the transverse vehicle axis B-B 'corresponding to the X axis, inclined in the sensor field 5 is arranged. The weight means 16 changes the same test signal t in such a way that a defined weighted output signal a _{g is} generated.

A first test specification for the sensor arrangement 5 provides a paired test of the sensors in such a way that an unweighted output signal of either the sensor sensing in the X direction 17 or the sensor that senses in the Y direction 18 together with the weighted output signal of the so-called safing sensor 19 to be considered together. In particular, the seismic mass 12 of the sensor element 11 of the first acceleration sensor 19 deflected with weighted force in the opposite direction of their sensitivity axis w or a corresponding signal a _wg electronically generated. In addition, the seismic mass 12 of the sensor element 11 of the second acceleration sensor 17 or 18 deflected with unweighted force in the direction of their sensitivity axis x or y or a corresponding signal a _x or a _y electronically generated. Of course, a reverse process is also conceivable, ie the generation of weighted output signal a _x or a _y in x or y and an unweighted signal a _w relating to the w direction.

In the former case, the weighting means 16 preferably change the test signal t such that the electrical output signal a _wg is output weighted by the mathematically expressed weighting factor k _w equal to ½ · √2 (which corresponds to approximately 0.707). In the case of sensors arranged differently inclined relative to one another, the weighting factor must be adapted accordingly depending on the angular position.

After a second test specification, all output signals in the sensor field 5 arranged sensors 17 . 18 . 19 common at the same time subject to careful consideration. In particular, the seismic mass 12 of the sensor element 11 of the first acceleration sensor 19 deflected with weighted force in the opposite direction of their sensitivity axis w or a corresponding signal a _wg electronically generated. In addition, the seismic masses 12 of the sensor elements 11 of the second acceleration sensor 17 . 18 deflected with unweighted force in the direction of their sensitivity axis x and y or a corresponding signal a _x and x _y electronically generated. In this case the weighting means 16 preferably change the test signal t such that the electrical output signal a _w is output a _wg weighted by the mathematically expressed factor k _w equal to √2 (which corresponds to approximately 1.41). Conceivable alternatives are again expressly included.

6 shows an alternative preferred sensor arrangement according to the invention. Contrary to the arrangement according to 5 is in place of individual sensors that sense in the X and Y directions 17 respectively. 18 a so-called XY sensor 20 intended. The sensitivity axis v of the safing sensor 19 is in turn inclined to these axes, in the present case by 45 ° to the vehicle longitudinal axis A-A 'corresponding to the Y axis. In this arrangement, too, the functionality of the control circuit can be carried out according to the two text instructions described above 2 being checked.

7 finally shows another preferred sensor arrangement according to the invention. Contrary to the arrangement according to 5 are of the at least three individual sensors that sense u, v or w along a sensitivity axis 17 . 18 . 19 additionally a second sensor 18 arranged inclined to the vehicle longitudinal axis AA 'or transverse axis BB', so that the acceleration sensors 17 . 18 . 19 each have differently aligned sensitivity axes u, v, w. The sensors are preferably 17 . 18 . 19 star-shaped, each offset by 120 °. A functional test of this arrangement is carried out according to a third test specification, that is, all in the sensor field 5 arranged sensors 17 . 18 . 19 be subjected to a simultaneous joint examination as described for the second test specification, in which case the weighting means 16 the test signal t against the safing sensor 19 has to be changed such that the electrical output signal is output weighted by the mathematically expressed factor k _W equal to 2.

Weighted and unweighted output signals from the acceleration sensors 17 . 18 . 19 . 20 are in the microcontroller 3 processed using a safing algorithm, which in turn forms a weighted sum Σ _g . If the processing yields a predetermined value, for example approximately zero, this is a sign that the safing algorithm itself also works reliably. If the predetermined value is to be approximately zero, the factor k _w according to the invention should preferably be selected such that the angular positions between the sensors are ultimately compensated.

The present invention thus advantageously allows not only the functional check of the G cell 11 and / or their control circuit 15 an acceleration sensor 17 . 18 . 19 . 20 , but also to check the functionality of a Safing algorithm that is subordinate to the Chrash algorithm. The present invention is therefore particularly suitable for occupant protection systems in a modern motor vehicle.

Claims (23)

Control arrangement ( 2 ) for occupant protection means in a motor vehicle ( 1 ), - whereby the control arrangement ( 2 ) a sensor field ( 5 ) with at least two acceleration sensors ( 17 . 18 . 19 respectively. 19 . 20 ) which is assigned at least two sensor elements (Gcells 11 ) which allow acceleration sensing along three sensitivity axes (u, v, w and w, x, y); - The sensitivity axes (u, v, w or w, x, y) of the sensor elements ( 11 ) of the acceleration sensors ( 17 . 18 . 19 respectively. 19 . 20 ) span a level, which after installing the control arrangement ( 2 ) in a motor vehicle ( 1 ) runs essentially parallel to a plane defined by a vehicle longitudinal axis (A-A ') and a vehicle transverse axis (B-B'); - with at least one evaluation device ( 3 ) Comprehensive - for normal or crash operation - a safing routine to check the plausibility of all output signals (a _u , a _v , a _w or a _w , a _x , a _y ) of the sensors ( 17 . 18 . 19 respectively. 19 . 20 ) by forming a weighted sum (Σ _g ) from the output signals (a _u , a _v , a _w or respectively a _w , a _x , a _y ), and - a crash routine for evaluating the output signals (a _u , a _v , a _w or respectively a _w , a _x , a _y ); and - for a test operation - a test routine which sends a test signal (t) to at least two acceleration sensors ( 17 . 18 . 19 respectively. 19 . 20 ) sends the purpose of generating output signals (a _u , a _v , a _w or respectively a _w , a _x , a _y ) to check the function of the sensors ( 17 . 18 . 19 respectively. 19 . 20 ); characterized in that - by means of a weighting means ( 16 ) at least one test signal (t) can be modified by a predeterminable weighting factor (k _w ) such that at least one acceleration sensor ( 19 ) outputs a weighted output signal (a _wg ); and - that during the test routine the output signals (a _u , a _v , a _wg or respectively a _wg , a _x , a _y ) of the in the sensor field ( 5 ) arranged acceleration sensors ( 17 . 18 . 19 respectively. 19 . 20 ) according to the Safing Rou tine can be processed, - whereby the weighted sum (Σ _g ) of the output signals (a _u , a _v , a _wg or respectively a _wg , a _x , a _y ) in the case of functional acceleration sensors ( 17 . 18 . 19 respectively. 19 . 20 ) should result in a predetermined value; and - whereby the control arrangement ( 2 ) can be determined if the weighted sum (Σ _g ) of the output signals actually delivered during the test routine (a _u , a _v , a _wg or a _wg , a _x , a _y ) gives approximately the predetermined value. Control arrangement ( 2 ) according to claim 1, characterized by a sensor field ( 5 ) with three acceleration sensors ( 17 . 18 . 19 ) comprising one sensor element each ( 11 ) for each direction of sensitivity (u, v, w or w, x, y). Control arrangement according to Claim 1, characterized by a sensor field ( 5 ) with a first acceleration sensor ( 19 ) with a sensor element ( 11 ) for a predetermined sensitivity direction (w) and by a second acceleration sensor with two sensor elements ( 11 ) for two predetermined sensitivity directions (x, y). Control arrangement according to claim 1, characterized by a first acceleration sensor ( 19 ) with a sensor element ( 11 ) for a predetermined sensitivity direction (w) and by a second acceleration sensor ( 20 ) with a sensor element ( 11 ) for two predetermined sensitivity directions (x, y). Control arrangement ( 2 ) according to one of the preceding claims, characterized in that the weighting means ( 16 ) Part of the evaluation device ( 3 ) and / or part of the acceleration sensor (s) ( 17 . 18 . 19 . 20 ) is. Control arrangement ( 2 ) according to one of the preceding claims, characterized in that the weighting means ( 16 ) comprises several so-called test fingers, a voltage-reducing component such as a resistor or a voltage-increasing component such as a charge pump or another suitable electronic and / or mechanical component. Procedure for checking the function of a control arrangement ( 2 ) for occupant protection devices ( 7 . 8th . 9 . 10 ) in a motor vehicle ( 1 ), in particular a control arrangement ( 2 ) according to one of the preceding claims, - wherein the control arrangement ( 2 ) a sensor field ( 5 ) with at least two acceleration sensors ( 17 . 18 . 19 respectively. 19 . 20 ) which is assigned at least two sensor elements (Gcells 11 ) which allow acceleration sensing along three sensitivity axes (u, v, w and w, x, y); - The sensitivity axes (u, v, w or w, x, y) of the sensor elements ( 11 ) of the acceleration sensors ( 17 . 18 . 19 respectively. 19 . 20 ) span a level, which after installing the control arrangement ( 2 ) in a motor vehicle ( 1 ) is substantially parallel to a plane defined by a vehicle longitudinal axis (A-A ') and a vehicle transverse axis (B-B'); - the control arrangement ( 2 ) at least one evaluation device ( 3 ), which - in normal or crash operation - all output signals (a _u , a _v , a _w or respectively a _w , a _x , a _y ) of the sensors ( 17 . 18 . 19 respectively. 19 . 20 ) checked for plausibility by means of a safing algorithm by forming a weighted sum (Σ _g ) from the output signals (a _u , a _v , a _w or respectively a _w , a _x , a _y ), and - the output signals ( a _u , a _v , a _w or respectively a _w , a _x , a _y ) is evaluated using a crash differentiation algorithm; and - in a test mode - a test signal (t) at at least two acceleration sensors ( 17 . 18 . 19 respectively. 19 . 20 ) sends for the purpose of generating output signals (a _u , a _v , a _w or a _w , a _x , a _y ) to check the function of the sensors ( 17 . 18 . 19 respectively. 19 . 20 ); characterized in that - at least one test signal (t) experiences a weighting (k _w ) such that at least one acceleration sensor ( 19 ) outputs a weighted output signal (a _wg ); and - that in test mode the output signals (a _u , a _v , a _wg or a _wg , a _x , a _y ) that in the sensor field ( 5 ) arranged acceleration sensors ( 17 . 18 . 19 respectively. 19 . 20 ) are processed in accordance with the safing algorithm, - the weighted sum (Σ _g ) of the output signals (a _u , a _v , a _wg or a _wg , a _x , a _y ) in the case of functional acceleration sensors ( 17 . 18 . 19 respectively. 19 . 20 ) should result in a predetermined value; and - whereby the control arrangement ( 2 ) is determined when the weighted sum (Σ _g ) of the output signals actually delivered during the test operation (a _u , a _v , a _wg or a _wg , a _x , a _y ) gives approximately the predetermined value. Method according to Claim 7, characterized in that at least one of the output signals (a _u , a _v , a _w or a _w , a _x , a _y ) is compared with a threshold value (SW), only when the threshold value (SW ) the safing algorithm is enabled by at least one of the output signals (a _u , a _v , a _w or a _w , a _x , a _y ). A method according to claim 7 or 8, characterized in that the test signal (t) of the control circuit ( 15 ) of the sensor ( 17 . 18 . 19 . 20 ) is supplied in such a way that an output signal (a _u , a _v , a _wg or a _wg , a _x , a _y ) is generated or simulated electronically. A method according to claim 7 or 8, characterized in that the test signal (t) the sensor element (Gcell 11 ) of the sensor ( 17 . 18 . 19 . 20 ) is supplied in such a way that the seismic mass ( 12 ) of the sensor element ( 11 ) is deflected in a predetermined direction (u, v, w, x, y). Method according to one of claims 7 to 10, characterized by the use of a so-called safing sensor ( 19 ) with a sensor element ( 11 ), whose sensitivity axis (w) is arranged obliquely to two perpendicular sensitivity axes (x, y), especially inclined by an angle of 45 °, 135 ° or 225 ° to the perpendicular sensitivity axes (x or y). A method according to claim 11, characterized by the use of two sensors ( 17 . 18 ) comprising one sensor element each ( 11 ) with a perpendicular sensitivity axis (x, y). Method according to claim 11, characterized by the use of a so-called XY sensor comprising two sensor elements ( 11 ) with one perpendicular sensitivity axis (x, y). A method according to claim 11, characterized by the use of a so-called XY sensor ( 20 ) comprising a sensor element ( 11 ) with two mutually perpendicular sensitivity axes (x, y). Method according to claim 9 or 10 and one of claims 11 to 14, characterized in that the seismic masses ( 12 ) two sensor elements ( 11 ) deflected in a predetermined direction or corresponding signals are generated or simulated electronically, in particular - that the seismic mass ( 12 ) of the sensor element ( 11 ) of a first acceleration sensor ( 19 ) deflected with a weighted force in the opposite direction of its sensitivity axis (w) or a corresponding signal (a _wg ) is generated or simulated electronically, and - that the seismic mass ( 12 ) of the sensor element ( 11 ) of a second acceleration sensor ( 17 . 18 ; 20 ) deflected with unweighted force in the direction of its sensitivity axis (x or y) or a corresponding signal (a _x or a _y ) is generated or simulated electronically; - or the other way around. A method according to claim 15, characterized by a weighting factor (k _w ) of one half of the root of two (½ · √2 ≈ 0.707). Method according to claim 9 or 10 and one of claims 11 to 14, characterized in that the seismic masses ( 12 ) three sensor elements ( 11 ) deflected in a predetermined direction or corresponding signals are generated or simulated electronically, in particular - that the seismic mass ( 12 ) of the sensor element ( 11 ) of a first acceleration sensor ( 19 ) deflected with weighted force in the opposite direction of its sensitivity axis (w) or a corresponding signal (a _wg ) is generated or simulated electronically; - that the seismic mass ( 12 ) of the sensor element ( 11 ) of a second acceleration sensor ( 17 . 18 ; 20 ) deflected with unweighted force in the direction of its sensitivity axis (x or y; u) or a corresponding signal (a _x or a _y ; a _u ) is generated or simulated electronically; and - that the seismic mass ( 12 ) of the second or a third sensor element ( 11 ) of the acceleration sensors ( 11 . 12 . 13 . 14 ) deflected with unweighted force in the direction of its sensitivity axis (y or x; v) or a corresponding signal (a _y or a _x ; a _v ) is generated or simulated electronically; - or the other way around. A method according to claim 17, characterized by a weighting factor (k) from the root of two (√2 ≈ 1.41). Method according to one of claims 7 to 10, characterized by a star-shaped arrangement of three sensors ( 17 . 18 . 19 ), each comprising a sensor element ( 11 ) with inclined sensitivity axes (u, v, w), especially with one sensitivity axis (u, v, w) at an angle of 120 ° to each other. Method according to claim 9 or 10 and 19, characterized in that the seismic masses ( 12 ) three sensor elements ( 11 ) are deflected in a predetermined direction or corresponding signals are generated or simulated electronically, in particular according to the features of claim 17. A method according to claim 20, characterized by a weighting factor of 2. Method according to one of the preceding method claims, characterized in that the weighted sum (Σ _g ) of the output signals (a _u , a _v , a _wg or a _wg , a _x , a _y ) must be approximately zero in order to be in test mode perfect functioning of the control arrangement ( 17 . 18 . 19 respectively. 19 . 20 ) to diagnose. Method according to one of claims 10 to 22, characterized by a capacitive test deflection of the seismic mass ( 12 ) of the acceleration sensors ( 17 . 18 . 19 respectively. 19 . 20 ).